1. Field of the Invention
The present invention relates to an optical waveguide device used for an optical communication system, and in particular, to a structure of an optical waveguide effective for the miniaturization of an optical circuit for monitoring the optical output power.
2. Description of the Related Art
An optical waveguide device is a device realizing various functions using an optical waveguide which confines a light within a high refractive index portion formed in a dielectric medium to propagate the light. For example, an optical waveguide device, which constitutes a Mach-Zehnder interferometer using a dielectric material, such as lithium niobate (LiNbO3: to be referred to as LN hereunder), is popularly used as an optical modulator, an optical switch, a variable optical attenuator or the like, since it has an extremely high electro-optic constant and has a response speed higher than that of a device having a thermal optic (TO) effect.
However, there has been known that the above described optical waveguide device using the dielectric substrate of LN or the like has the potentiality of occurrence of a phenomenon called the temperature drift in which an operating point is shifted due to a temperature change or a phenomenon called the DC drift in which the operating point is shifted by applying a direct current. When the operating point is shifted due to the occurrence of the temperature drift or the DC drift, an optical output characteristic of the optical waveguide device are fluctuated. Therefore, in the optical modulator for example, the modulation cannot be performed under a constantly fixed condition.
To be specific, the optical output power of the Mach-Zehnder optical modulator is changed in accordance with cos2(Δφ/2). The above described parameter Δφ is a phase-change amount given by an interaction section in the Mach-Zehnder interferometer, and is represented by a relationship of Δφ={π·ne3·γ33·l/(λ·d)}·V in the case where the LN substrate of Z-cut is used. In this relationship, ne is a refractive index of the optical waveguide, γ33 is an electro-optic constant, l is the length of each of electrodes disposed on two parallel optical waveguides, λ is an optical wavelength, d is a distance between the electrodes, and V is an applied voltage. The optical output power characteristic of the optical modulator is represented by a curve as shown in FIG. 19, provided that the horizontal axis indicates the applied voltage V.
In the optical modulator as described above, it is desired that the operating point is set to be in an intermediate state between the ON and the OFF when the applied voltage to the electrodes is 0V. However, an actual operating point is often deviated from a desired operating point due to a manufacturing error or various stresses. In order to solve this deviation of the operating point, it is typical that the operating point is adjusted to the desired operating point by applying a direct-current bias voltage. However, the operating point adjusted by the DC bias voltage is shifted due to the above described DC drift. Therefore, in order to stably realize the desired operating point, it is necessary to always monitor the optical output power to control the DC bias voltage based on the monitoring result. This monitoring of the optical output power is not limited to the case where the optical waveguide device is used as the optical modulator, and is necessary to adjust an optical attenuation amount corresponding to the temperature change or the like, in the case where the optical waveguide device is used as the variable optical attenuator of Mach-Zehnder type for example.
For coping with the above described necessity of the optical output power monitoring, heretofore, there has been proposed a technology in which a light receiving element for monitoring the optical output power is disposed in the optical waveguide device (refer to literature 1: Japanese Unexamined Patent Publication No. 2002-182050).
As the conventional optical waveguide device, there has been known the device of a configuration in which a main signal light emitted from an end face of the optical waveguide is led to an output optical fiber via a lens coupling system or the device of butt joint type in which the end face of the optical waveguide is directly abutted with the output optical fiber. In the configuration using the lens coupling system, since there exists a required space between a substrate side face from which the main signal light is output and the lens coupling system, of the optical waveguide device, the light receiving element for monitoring the optical output power can be arranged by utilizing this space, and accordingly, it is possible to receive a sufficient monitor light.
On the other hand, in the case of the device of butt joint type, since the output optical fiber is extremely thin, the intensity thereof is insufficient if the fiber is simply adhered to the end face of the optical waveguide. Therefore, as shown in (A) of FIG. 20 for example, it is necessary to fix an output optical fiber 110 to an end face of an optical waveguide 101A on the main signal light output side, using a fiber fixing member 120 such as a V-groove fiber block, glass ferrule or the like. In the configuration using this fiber fixing member 120, if a light receiving element 130 for monitoring the optical output power is arranged on the rear side (opposite side to the optical waveguide device) of the fiber fixing member 120, the fiber fixing member 120 hinders the light receiving element 130 from sufficiently receiving a monitor light emitted from an optical waveguide 101B on the monitoring side. In order to avoid such a situation, as shown in (B) of FIG. 20 for example, the fiber fixing member 120 needs to be formed in a complicated shape. The reinforcing capillary shown in the literature 1 is considered to be one example of the fiber fixing member of more complicated shape. The complexity of the fiber fixing member has a problem of the cost rise of the optical waveguide device.
As one measure for solving the problem in the above butt joint type configuration to reliably monitor the optical output power, it is effective to lead out the monitor light from a substrate side face (side face positioned on the front side or the back side in the configuration specifically shown in FIG. 20) different from the substrate side face from which the main signal light of the optical waveguide device is output. However, in order to realize such a configuration, it is necessary to solve the following problems.
As shown in FIG. 21 for example, a first problem is the reflection and radiation loss on the substrate side face in the case where the monitor light is led out using a curved waveguide. Namely, considering the LN modulator as a specific example, the width w of a substrate 100 of the normally used LN modulator is about 1 mm to 2 mm. Therefore, in order that the light propagated through a curved waveguide 101B on the monitoring side is led out at an angle at which the light is not totally reflected by the substrate side face, the curvature radius Rc of the curved waveguide 101B needs to be set to around 1 mm to 2 mm. On the other hand, the curvature radius Rc of 30 mm or above is necessary to avoid an occurrence of the radiation loss in the curved waveguide 101B. Therefore, as shown in (A) of FIG. 21, if the curvature radius Rc of 30 mm or above is ensured in order to avoid the radiation loss in the curved waveguide 101B, the monitor light is totally reflected by the substrate side face, and also the substrate size is not made to be larger. Further, as shown in (B) of FIG. 21, if the curvature radius Rc of the curved waveguide 101B is set to 2 mm or less in order to prevent the total reflection by the substrate side face, the monitor light is radiated to the outside of the waveguide, in the halfway of the curved waveguide 101B. Accordingly, it is hard to receive the sufficient monitor light only by simply forming the curved waveguide.
As shown in FIG. 22 for example, a second problem is the difficulty of receiving the monitor light caused by the chipping which is generated on the surface of the substrate. Namely, a chip forming the LN modulator or the like is obtained by cutting out the substrate material by utilizing a dicing apparatus. However, when the chip is cut out, the irregularity of several ten μm is generated on the top surface or the bottom surface of the chip. This irregularity is called the chipping. The LN modulator chip is provided with the optical waveguide which is formed on the top surface of the chip by the diffusion treatment of Ti or the like. Therefore, if the chipping is generated on the substrate side face from which the monitor light is led out, it becomes difficult to obtain the sufficient monitor light. Accordingly, it is necessary to take the countermeasure against the chipping on the substrate side face from which the monitor light is extracted.
A third problem is the difficulty of reliably mounting the light receiving element for receiving the monitor light. Namely, as one of methods of mounting the light receiving element for receiving the monitor light, there is considered a method of attaching the light receiving element to the substrate side face from which the monitor light is led out. However, since the size of the light receiving element is 300 μm or above, if the light receiving element is attached to the substrate side face in the case where the optical waveguide is formed on the upper portion of the substrate as described in the above, the light receiving element runs off the top surface of the chip as shown in FIG. 23. Therefore, it becomes extremely difficult to attach the light receiving element, and consequently, problems inclusive of the reliability occur.